Win Payout Frequency Graph
 

This is a frequency distribution graph for Win Payouts based on my modest dirt sprint database (171 races). For The Lead recommends that you dismiss all horses with M/L Odds of 20-1 or greater. That's good guidance. And as you can see from the graph, there aren't many winners beyond 10-1 P/T Odds. So anytime you bag one in that territory, take pride. It's no mean accomplishment.

http://i66.tinypic.com/11ah380.jpg

Hi Mick
 

May I ask a question,
How were the 171 races compiled,

Were these your most recent 171 races? Are these races that you handicapped before the results were known ? What percentage of these 171 race did you actually place a bet ?

Did you use a predetermined pace line / contender selection method ?

There probably is a second part to my question
Look forearm to your reply
Good Skill
Bill

Hi, Bill.

These are recent dirt sprints (all 2017) from the tracks I routinely play, e.g. Parx, MVR, OP. There was really no handicapping involved for purposes of my statistics. I used the "export" function in RDSS, pasted the winners' lines into a modified version of the MasterTemplate3, added a few extra values, and then imported the records into Access. How many did I bet? Maybe 20%. I didn't keep track but I usually bet maybe 2 races per card. And with very few exceptions, I used the default paceline.

Thank You
 

Thank you Mick

I'm doing my own research, Not so much in how often things happen
withing a test cycle , but rather the way people record the data,

I appreciate your reply and interesting work

Good Skill
Bill

One other point about the graph, if I had broken the ranges down further, e.g. 2-1 and 5-2, 3-1 and 7-2, 4-1 and 9-2, the three tall sticks would have been shorter, but I think they would have still been the tallest six.

(Oops. Our messages crossed in cyberspace. And that's not what you're interested in. (-: )

With more granularity, our frequency distribution looks like this:

http://i63.tinypic.com/28u4oxu.jpg

Over 20/1 can make very good Place and Show bets. I generally toss them out of the mix for win but I do look at them to place or Show. I hit one the other day that payed me 17.00 for show, the horse was 62/1 at post. This is, of course, rare, but I have hit several 20+/1 horses that paid very well in the lesser straight bets. These horses will have corollaries that support a wager so I give each horse a line and then keep an eye on them at post time.

20
 

Good post Fairchild ,

Along with place and show, sometimes I can add a 20/1 in one leg of my pick 4
Often in one of the legs I can go 5 horses deep, in these races I might add it as its just a .50 cent unit or $1.

Higher priced winners
 

Mitch , in looking over your graph I see that 55/171 winners (32%) paid 12.00 and up.
I have kept track of the 12.00 and up winners at OP this year.
In 320 races there has been 123 wins. this is 38% of all the playable races.
82 of the 123 were actually in my final contenders.
rmath

Rich,

That's a great point. We always focus on post-time favourites (Ted's spelling) winning one third of the races. Well, horses paying $12+ (5-1 and greater) win a third of the races, too. At least they did in my dirt claimer database. And the percentage is even higher with your statistics. Also impressed that you identified them as contenders two thirds of the time. That's some good handicapping. Thanks for your insight.

Fairchild, as I'm a lousy Place and Show bettor, I use them underneath in Exactas. Chalk over Odds usually pays nicely. For example, Parx 8th race on Tuesday, 2-1 WT 28-1 paid $124 and the 28-1 horse was very "findable" with a recent second and third, if I recall correctly.

So take the next step
 

Now that you have a distribution of payouts, why not look at those larger payoffs and see if you can find the paceline that would have put the horse in your final contenders? What are the common features of these horses? If 38% of your winners paid $12 or more, isn't that as much or more than favorites or strongly bet down horses? Doesn't it make sense to develop a strategy to consistently wager on these longer priced horses.
If you study Bradshaw you will see how he did it. We are not reinventing the wheel here!

Mark, that is a good suggestion and I'm thinking about how to go about it. I'm probably not the right person for the specific approach you mentioned though because I seldom change a paceline. (Perhaps we're trying to solve the same puzzle using different methods.) But it would be good to know what factors are shared by the $12+ winners versus those of the lower-odds winners. A betting strategy could be developed from the differences.

BTW, I created these Frequency Distribution graphs while looking for a statistically orthodox method for trimming the extremes which skew the data. I knew there was a problem with my data set because the average payout was ~$12 while the harmonic mean was ~$7.

I first thought of the Empirical Rule (68% within one standard deviation, etc.) but I didn't have a normal distribution like the "bell curve" we studied in statistics. So standard deviations wouldn't work. (That's where the Frequency Distribution graphs came from.)

I next looked at Inter-Quartile Range and "outliers." That didn't work either because it only eliminates the high-odds horses and leaves all the odds-on chalk which you don't need a computer program to find.

I still don't have a satisfactory approach but while mulling this over, I thought I would post the graphs. I appreciate the comments. And if anyone has a suggestion about trimming the extremes, please let me hear from you, whether by post or PM.

Enjoy the races...
 

Mick,
I don't believe it is possible to use and automated paceline selection routine in handicapping a race and make any serious money, or any money really unless you are very, very selective in the races you play. But that's just my opinion. Our wagering here in the US is Para Mutuel and we are competing against the best handicappers in the world. Large syndicates spend more money on software development than we pay annually for our home mortgage. Again my belief is that you have to operate outside the mainstream, in other words, use a different strategy, one that is not reliant on regressing of public factors. The obsession with recent good races insures that any horse that exhibits a speed or pace advantage will be hammered at the windows.
I have spent the last 4 years studying the Hat Check blog and have had the good fortune to exchange emails with Richie P concerning Match Up concepts. One of the things that impressed me most about Jim Bradshaw's work was his notion of comparing horses. Jim's belief was that he was more interested in a horse's best effort than his last one. Think about it. I want to know the fastest pace this horse ever competed against and won or finished up close. If you compare horses this way you know who is, was, the most talented horse entered in today's race. In other words, if he runs that race today, he wins. Now if that paceline is way down at the bottom of his pps, you have a dilemma. What is this horse now?
Is it not a true statement that today's winner will have had to compete successfully against the projected 1st fraction velocity? You may have horses that can go 21.8 but your automated paceline selection routine picks lines based upon Total Energy and the line picked for some horses could be 22.4, 23.2 or 24 seconds. A horse that can run well and finish against a 24 second 1st fraction is going to be way back today. If he attempts to run faster earlier chances are that this exertion will nullify and late gain. At the 2nd call this horse won't be anywhere near contention. So to handicap this way defies all laws of physics. Fast horses win races. Fast horses that run unchallenged on the lead win alot of races. A thorough examination of the horse's pps will tell you just about everything you need to know about a horse.
If you did nothing more than find the horse in today's race that runs Early and you expect to get the lead out of the gate. Then using that time as translated to fps velocity on the Velocity-POR screen, find the lines for all the other horses that are equal or similar. Those will be your contenders. If a horse has never shown that he can finish against that 1st fraction velocity and you do not perceive a pace duel up front then that horse is not a contender. Do not use lines on S horses that ran against faster paces because if they do not get that same early pace they will not gain the ground late. Track this method! Find out when a deep paceline is not usable today. Lastly,
"THE Hat" gave 3 questions you must answer if you choose a deep paceline: 1) Is the horse still a horse? 2) Has he changed his running style and become a slow horse? 3) If he is Early can he still get on top of his fractions?
Handicapping the ponies can be a very enjoying hobby. It can be a diversion from the rigors of life. But the expectations of this type of involvement should be minimizing the amount of money you lose. Like going to Las Vegas with a set amount of money you can lose. Good Luck and have some fun!!

Thanks, Mark. I sincerely appreciate your thoughts and opinions. I have one of your posts stuck on the cork board over my computer. It has to do with TPR and the Match Up. You may remember it. It expresses several of the same ideas. I'll probably print this one and stick it up there, too. :)

Quantitative Analysis
 

Mick,
One last word, as I believe this is very important when looking at wagering on Thoroughbred Horse races. Someone along the line came up with the idea that you can accumulated races in a database and draw meaningful information from them. However, each race in an independent event, just like a coin flip. While they are not random events, identifying the causation and the effect is the difficult task. These are living breathing beings that are influenced by weather, surface speed, the wishes of their connections and the trips they experience running around the track. We can attempt to quantify these factors as it relates to the internal fractions, final time and the horse's exhibited behaviors in his past performances, but at best they are crude estimates.
We must remember that horses are creatures of habit. A visual Running Style is an expression of that habit. When a horse is placed in a race in which the early pace is within his ability he will run similarly. When placed against superior faster animals his behavior will be much different. If you accept the Running Style concept as valid then it follows that there are other aspects of his behavior that note his individuality. Need-to-lead characteristics, competitiveness or fighting characteristics, the ability to pass other horses and so forth. As well, a horse has to have enough work to attain a condition of racing fitness. Some combination of workouts and racing can bring a horse to his peak of conditioning without necessarily stacking a bunch of wins or close finishes in his recent races. If you look at Tam's first race today that was posted on this board, the winner came from the Fastest Pace Last Race and battled through 1/2 mile within a head of the leader before fading to last beaten 13 lengths in a 7f race at 43/1. Jim Bradshaw discussed Hidden Early and a paceline with the first two calls of 2nd-hd - 2nd-hd as an absolute assurance that a horse is approaching his peak condition. Today on a two level drop and 1f cutback he wins at 8/1. No amount of database work or statistical evaluation is going to put you on that horse. We can't really know beforehand how much energy it took out of the horse to battle for the lead for 1/2 mile. That is not his Running Style he has won twice as an SP and once as a P. Today he reverted back to his normal Running style and won fairly comfortably. What it does tell you, thank you Jim Bradshaw is that the horse is peaking and back in 17 days is sound down two levels in Claiming price cutting back, he is placed for action.
Sorry for my long winded response but we as humans want to create order out of the chaos that is horse racing and in most cases that is accumulating numbers and deriving factor lists that give us a sense of confidence in wagering our money. This ignores the information that is there for us to consider if we look beyond the numbers and recognize the general talent and capabilities of the animal and his fitness compared to what he will face today. Each race is a different combination of horses and their cumulative energies will determine who will prevail today.

mick,
question........the frequency #'s that you posted are winning bets that you made or just winner's frequency numbers?

frequency
 

mick,
question........the frequency #'s that you posted are winning bets that you made or just winner's frequency numbers?

They are winner frequency numbers, not winners that I bet.

a frequency of your bets and the roi attached to it would be more a realistic model of your
system, beside i'm a firm believer of mastering one track, where most of the same horses do race against each other. Good luck!

I apologize for any misunderstanding. I'm doing some statistical research for Ted on the ROIs of Rx2 and Rx3 compared to MLO# and PTO#. This isn't about my own system. (I don't actually have a system and bet sparingly, but have managed a 1.23 ROI for the year.)

I created the Frequency Distribution graph to determine if I could use the Empirical Rule and Standard Deviations to trim the extreme values that skew the data and distort results. I wanted a more accurate statistical picture, so to speak. Once I had compiled it (and realized I didn't have a Normal Distribution), I thought it might be of interest to others.

It's only a graphical representation of the win payout frequencies of the dirt sprints (mostly claimers) in my database. Hopefully, some do find it interesting. :)

see if this can help you......"Say you have a sample of 100 races. If the mutuel prices are sorted from lowest to highest, the lowest may be 2.20 and the highest
89.00. The purpose of modeling is predicting the future.
That purpose can best be fulfilled by using an approximation that is far more likely to be repeated
than simple averaging of the 100 mutuel prices would indicate.

Using the list of sorted mutuel prices, pick the 25th from the bottom and the 25th from the top.
Those values may be 8.00 and 4.00. Add to get 12.00. Divide by 2 to get 6.00--which is an accurate
predictive value of what future average prices will be.

Multiply the 6.00 by 1.5 to get 9.00. Set that as a default value to replace
any mutuel price greater than that amount and go through the list of mutuel prices again using:

If (mutuel price) > 9.00 then (mutuel price) = 9.00

Use the new cleaned data to calculate an ROI.
If it is positive, you may have something that can usefully be applied to future races.
Not guaranteed, but a much more dependable model than using dirty data...."

BTW, the MIQR is used routinely in predictive computer modeling and simulations. It is not some personal quirk of mine, good luck.

gl45, thank you for your suggestion. The "M" in MIQR stands for ... ?

M=modeling....you already know what IQR stand for.

Thank you, gl45. I enjoy learning new things. :)

mick,
one factor frequency, MLO or PTO, about a bunch of factors. Here is a good reading.

http://racingtopics.blogspot.ca/2017...g-matters.html

Thanks, gl45. That's a fine article. (Did you write it?) It captures much of my thinking about playing the horses.

mick,
are you kidding.....me writing that kind of article about racing handicapping.
Good luck at the races

Hi, gl45.

Since we all use aliases, I wasn't sure if I was corresponding with Jim Quinn. ;)

In my world PTO is Paid Time Off. :)

Just curious if any of you math geniuses have ever thought about constructing a "quality" metric, in the sense of what objective advantage (for example) a 2 horse dutched selection has over the rest of the field, or the next three (or all) of your rated runners ? I'm trying out a few ideas. For example today (7/22) at Woodbine, in the 4th race, my "Quality Metric" had the # 6 with a higher than normal advantage over the next three, (my 5 total contenders). If this sounds like something anyone has worked on, I'd be interested in any ideas you have.

Just saw this post Mick and some how missed it. Your payout study basically mirrors a study done many years back by Burton Fabricand and published in1965.

What his statistical study found was that horses win according to their odds, or in other words the greater the odds the fewer the frequency of win. Booking tables were developed based on the percentages of wins and still hold up today.

I believe lumping them all together is a fools errand. Essentially if one wants to know about the unknown than study the unknown.

Mitch 44